LED lights

  • Active since 1995, Hearth.com is THE place on the internet for free information and advice about wood stoves, pellet stoves and other energy saving equipment.

    We strive to provide opinions, articles, discussions and history related to Hearth Products and in a more general sense, energy issues.

    We promote the EFFICIENT, RESPONSIBLE, CLEAN and SAFE use of all fuels, whether renewable or fossil.
Isn't Germany on 230VAC lighting, if I recall?

I have interest in switching 20 of these 6" cans from incandescent to LED, both for efficiency, and hopefully lifetime. My wife and kids tend to leave these 20 lights on all day and all night, I'm forever turning them off when I pass through, but then find them back on the next time I come around. It's the kitchen and great room, so usage is high, no matter what.

Some require a 10 foot ladder to reach, so I'd love to get something with longest possible life, not that I've actually ever seen longer life out of LED's vs. incandescent play out in reality. Also, burning 20 x 45W incandescent bulbs 8-12 hours per day has a way of increasing air conditioning usage. I bought incandescents with the highest efficiency I could find (370 lumens / 45 watts), but it's still like running a 900 watt heater in that space all afternoon and evening, in July.

I just can't seem to find bulbs that will work. I would even drop down to BR30 for the kitchen, the cans are really shallow, so a short-neck bulb is really needed to not protrude. I just don't want to see the back-metal of the can, if the bulb doesn't mostly fill the surround.
I've used compact florescent bulbs in this situation. Don't know if they make those anymore. They are not to far off LED's heatwise and efficiency wise as well. The color temp is right on.
 
I appreciate your perspective, and I agree on all points, except I'd also counter by saying manufacturers need to design these things to not fail when installed in the most common legacy fixtures. It's just good business. If the bulb is marketed as suitable for halogen replacement use, then it simply must operate reliably in fixtures designed for halogen bulbs!

These cans have ceramic sockets spring-clipped into the aluminum can. I'm not sure what the KT of the ceramic is, but there should be a thermal conduction path out through the wiring that's not totally terrible. The primary bottleneck would be the stranded (probably AWG-18?) pigtail on the fixture itself, and it's attachment point at the socket. If the ceramic has any KT above 1.5~2 W/m°K, then it'd also be sinking reasonable heat out to the aluminum shell of the fixture, but I don't really know anything about the type of porcelain or ceramic used in light fixtures.
In general you are correct, but only as long as government regulations and general customer expectations allow.

For the regulations, government can require they be manufactured (through various efficiency rules) so they have forced a design which won't work. We are suffering from some of this today. Short of changing regulations, you can't easily subvert this if customers have certain expectations around them.

For the general customer expectations, customers want cheap bulbs like the incandescents (or the halogens), they want them to work in all legacy fixtures, and they want them to look almost exactly like the filament based bulbs. What they don't realize is that's an unreasonable position in some fixtures and systems.

First, the heat transfer mechanism is completely different, and some fixtures won't accommodate that. Your overhead cans (IC or NIC) are not designed to move heat by convection but by radiation. And free/natural convection (no fans or external air movers) can't push hot air down and out of the cans. Most LED bulb failures in cans are due to heat problems (as long as water intrusion or moisture are not an issue). So if the cans are in a hot location that is higher than the bulb rating, it'll fail.

Secondly, if you can come up with a design that works in the legacy fixture, it often won't look like the legacy bulb. Consumers often shy away from those and they do not sell as well. While working at Rambus, I designed this bulb which had some higher heat transfer performance, and optics that matched an incandescent bulb:
3104cc40deab306b3bd5b19cf875d7d4a41fc066-1.jpeg


Due to the unique cylindrical optics, I accidentally discovered a method to move air horizontally in the center cylinder and cool it for horizontal applications. It also worked well vertically.

LED lightsLED lights
This bulb design (and variants) were sold in Amazon for a while.

The complaint? It didn't look like a light bulb. Working better and lasting longer didn't matter to the consumer.

While this bulb worked in legacy lamps, the real trouble is trying to use it in a legacy can. The various PAR, BR, etc., shaped bulbs do work in cans but moving air enough to cool an LED bulb without fan isn't feasible. I created a number of designs and tried them (some physical tests, some CFD simulations like those shown above) but nothing worked. The wires you mentioned might work to carry the heat but you can't get the heat to them easily through an Edison base/socket. The thermal contact resistance in those bases is too high and too variable. Conduction is often killed by the contact resistances, even if the thermal conductivity of the base material is reasonable.

Fan designs for NIC or IC fixtures would work, but consumer resistance was extremely high. Noise is also a concern, especially in thin sheet metal assemblies as they may vibrate with the fan frequency. I did make a can LED light that used special piezoceramic fans (little flapping metal shims, if you will) which were ultra quiet and highly efficient. That was installed in a company demo room and ran for years. But - the heat sink and fans were on top of the can, and this could not be used in an IC installation.

LED lights

So, bottom line, what you ask for isn't feasible (legacy sockets and heat transfer), and isn't wanted by the average consumer (they won't pay for it), and may have features they won't like (the appearance). You have to consider yourself a small segment of the market and that means finding your own solution, in part. What you want isn't fitting into the large marketplace and the companies aren't going to chase the 0.01% of the market looking for it. Personally in your situation I'd be making can modifications for better cooling.

Several years ago at GE I helped design an under cabinet "hockey puck" light around a unique form factor LED module, the GE Vio. The color performance was unmatched in its day as it used a near UV LED (405 nm) instead of the typical 450 nm blue. Special phosphors provided a rich color spectrum which enhanced looking at any food during preparation. It was thicker than a typical puck for the thermal design I made (about 1"). But today, almost 20 years later, they are working fine, and are on 12 hours or more a day. I've only replaced the external power supplies once. If you do the design right, it works - and ignore at least one limitation (appearance in this case). And none of the cheaper under cabinet LED lights today beat them, except on thinness.
 
In general you are correct, but only as long as government regulations and general customer expectations allow.

For the regulations, government can require they be manufactured (through various efficiency rules) so they have forced a design which won't work. We are suffering from some of this today. Short of changing regulations, you can't easily subvert this if customers have certain expectations around them.

For the general customer expectations, customers want cheap bulbs like the incandescents (or the halogens), they want them to work in all legacy fixtures, and they want them to look almost exactly like the filament based bulbs. What they don't realize is that's an unreasonable position in some fixtures and systems.

First, the heat transfer mechanism is completely different, and some fixtures won't accommodate that. Your overhead cans (IC or NIC) are not designed to move heat by convection but by radiation. And free/natural convection (no fans or external air movers) can't push hot air down and out of the cans. Most LED bulb failures in cans are due to heat problems (as long as water intrusion or moisture are not an issue). So if the cans are in a hot location that is higher than the bulb rating, it'll fail.

Secondly, if you can come up with a design that works in the legacy fixture, it often won't look like the legacy bulb. Consumers often shy away from those and they do not sell as well. While working at Rambus, I designed this bulb which had some higher heat transfer performance, and optics that matched an incandescent bulb:
View attachment 327773

Due to the unique cylindrical optics, I accidentally discovered a method to move air horizontally in the center cylinder and cool it for horizontal applications. It also worked well vertically.

View attachment 327774View attachment 327775
This bulb design (and variants) were sold in Amazon for a while.

The complaint? It didn't look like a light bulb. Working better and lasting longer didn't matter to the consumer.

While this bulb worked in legacy lamps, the real trouble is trying to use it in a legacy can. The various PAR, BR, etc., shaped bulbs do work in cans but moving air enough to cool an LED bulb without fan isn't feasible. I created a number of designs and tried them (some physical tests, some CFD simulations like those shown above) but nothing worked. The wires you mentioned might work to carry the heat but you can't get the heat to them easily through an Edison base/socket. The thermal contact resistance in those bases is too high and too variable. Conduction is often killed by the contact resistances, even if the thermal conductivity of the base material is reasonable.

Fan designs for NIC or IC fixtures would work, but consumer resistance was extremely high. Noise is also a concern, especially in thin sheet metal assemblies as they may vibrate with the fan frequency. I did make a can LED light that used special piezoceramic fans (little flapping metal shims, if you will) which were ultra quiet and highly efficient. That was installed in a company demo room and ran for years. But - the heat sink and fans were on top of the can, and this could not be used in an IC installation.

View attachment 327776

So, bottom line, what you ask for isn't feasible (legacy sockets and heat transfer), and isn't wanted by the average consumer (they won't pay for it), and may have features they won't like (the appearance). You have to consider yourself a small segment of the market and that means finding your own solution, in part. What you want isn't fitting into the large marketplace and the companies aren't going to chase the 0.01% of the market looking for it. Personally in your situation I'd be making can modifications for better cooling.

Several years ago at GE I helped design an under cabinet "hockey puck" light around a unique form factor LED module, the GE Vio. The color performance was unmatched in its day as it used a near UV LED (405 nm) instead of the typical 450 nm blue. Special phosphors provided a rich color spectrum which enhanced looking at any food during preparation. It was thicker than a typical puck for the thermal design I made (about 1"). But today, almost 20 years later, they are working fine, and are on 12 hours or more a day. I've only replaced the external power supplies once. If you do the design right, it works - and ignore at least one limitation (appearance in this case). And none of the cheaper under cabinet LED lights today beat them, except on thinness.
All interesting, but completely irrelevant to the failures mentioned above, which were candelabra bulbs in lamp posts, not reflectors in cans. With a total wattage roughly 14% - 15% of the bulb for which the fixture was designed, and power dissipation far below that due to difference in net efficiency, claiming the failures were due to heat is a tough sell.

I understand the problems well, I spent most of my career cooling high power transistors and associated bias circuitry. I was on the high-dollar low-volume side of the business, where we did not have to worry nearly so much about budgetary constraints, so our solutions were very different. But the underlying problems were the same.
 
Last edited:

Yep, that's the closest I found, as well. It's a bit long, would hang out of my fixtures, but I think they might also make a BR30 variant of that bulb that's about an inch shorter. CRI is 90, whereas @woodgeek is saying I should stay north of 95. But I think some of the bulbs I was complaining about earlier might have been as poor as 82.

Like @stoveliker said, a bit brighter than I was aiming, but it might be the best compromise. Presently, we're running 20x 45W bulbs, and that's a good lighting level with the dimmer sliders maxed.

None of the parametric search tools I've found include CRI + length + size + lumens. In fact, CRI is hard to find at all, in many cases. Will probably need to build my own spreadsheet, if I really want to pursue this.
 
Last edited:
What's interesting is that high CRI is possible in a short neck, but somehow only available in narrow spot with nasty blue color temperature, such as this PAR38:


If someone would just widen that out to a flood (eg. BR40) and drop the color temp to 2600 - 2700°K, it'd be perfect. Brighter than I want, but I can deal with dimmers. However, it seems that CRI nosedives as soon as color temperature comes down.
 
  • Like
Reactions: stoveliker
Of course; that has to do with the fraction of red one wants in the full spectrum.
Low temp high CRI is relatively much more red than high temp high CRI
 
  • Like
Reactions: Ashful
Heh... found one reasonable option: GE 65423, which is PAR38, but with a half-reasonable 40° beam angle (versus many are 25° to 35°). It has CRI = 90 and an impressive 50k hour lifetime rating, in a shallow 5.31" deep housing, and 2700°K color temp. Basically almost everything I was looking for.

Except... it's been discontinued, with no equivalent replacement available. ;lol

Finding a few short-neck BR40's, which would be ideal, but nothing under "stare at the sun" luminosity.
 
  • Like
Reactions: peakbagger
Yep, that's the closest I found, as well. It's a bit long, would hang out of my fixtures, but I think they might also make a BR30 variant of that bulb that's about an inch shorter. CRI is 90, whereas @woodgeek is saying I should stay north of 95. But I think some of the bulbs I was complaining about earlier might have been as poor as 82.

Like @stoveliker said, a bit brighter than I was aiming, but it might be the best compromise. Presently, we're running 20x 45W bulbs, and that's a good lighting level with the dimmer sliders maxed.

None of the parametric search tools I've found include CRI + length + size + lumens. In fact, CRI is hard to find at all, in many cases. Will probably need to build my own spreadsheet, if I really want to pursue this.

90 is fine for most people in living areas and kitchens. Sometimes higher CRI is good in dining areas bc it makes your food 'pop' a bot more. Makeup and dressing areas also like higher CRI.

Have you looked at the can replacement units that pop in, with a pigtail that plugs into the socket, and a fitting that clips into the opening, leaving the can in place. You could get some from amazon, try them out and return if they sucked, and stock up when you found a winner.
 
  • Like
Reactions: Ashful
Almost pulled the trigger on a 10-pack of "CRI94" Sunco BR40's on Amazon, until I visited manufacturer's site and found Sunco does not make a BR40 above CRI = 80. Misleading titles in Amazon products really doesn't help this search.
 
90 is fine for most people in living areas and kitchens. Sometimes higher CRI is good in dining areas bc it makes your food 'pop' a bot more. Makeup and dressing areas also like higher CRI.
Thanks! 90 is the new target, as I'm having trouble even finding that high, in lower color temperatures and lumens.

Have you looked at the can replacement units that pop in, with a pigtail that plugs into the socket, and a fitting that clips into the opening, leaving the can in place. You could get some from amazon, try them out and return if they sucked, and stock up when you found a winner.
I have not, mostly because the search was already too length and complicated for just "simple" replacement bulbs. I can't imagine the can of worms I'd be opening if I opened up to can replacement units... but maybe I should.

Six of these units in the great room are in the sloped part of a vaulted ceiling, with beveled cans, which might make can replacement even more fun. I don't even want to think about how many in the kitchen are going to come out without taking some paint. Ten in the kitchen are set into a beaded plank wood ceiling, painted off-white Alkyd, impossible to color match after years of aging. :(
 
These are 92 CRI, and fit 5/6" cans.


They are 1000 lumens, so you would want to put in a LED-dimmer, and tone it down a bit. :eek:
 
I have not, mostly because the search was already too length and complicated for just "simple" replacement bulbs. I can't imagine the can of worms I'd be opening if I opened up to can replacement units... but maybe I should.

Six of these units in the great room are in the sloped part of a vaulted ceiling, with beveled cans, which might make can replacement even more fun. I don't even want to think about how many in the kitchen are going to come out without taking some paint. Ten in the kitchen are set into a beaded plank wood ceiling, painted off-white Alkyd, impossible to color match after years of aging. :(

I think the bulbs you were looking for are getting scarce, bc most people prefer the retrofit units with the pigtails to get a 'flush' look. Much easier than stocking a bunch of lengths and diameters, and handling all those returns.

Newer units will do different color temps (with a dip switch inside the unit) so that is nice if you change your mind.

My late FIL went to the hardware store to but LED bulbs, and got these v cheap 5000K bulbs 'on sale' and put them in. The bulbs were also super low CRI and blue. We were shocked, and he thought they looked fine! We then realized he had bad cataracts in both eyes and couldn't see blue light anymore.
 
Last edited:
Yep, that's the closest I found, as well. It's a bit long, would hang out of my fixtures, but I think they might also make a BR30 variant of that bulb that's about an inch shorter. CRI is 90, whereas @woodgeek is saying I should stay north of 95. But I think some of the bulbs I was complaining about earlier might have been as poor as 82.
Warm white is already going to bias color slightly. That said, we have 2700k Feits throughout the house and the lighting is quite natural looking and very good for the kitchen where you want the color rendering to be fairly accurate.
 
  • Like
Reactions: Ashful
I think the bulbs you were looking for are getting scarce, bc most people prefer the retrofit units with the pigtails to get a 'flush' look. Much easier than stocking a bunch of lengths and diameters, and handling all those returns.
Okay, I'm sold. Just started searching, and while I didn't see exactly what I want, already I can see the landscape is way better in retrofits than BR bulbs. Wouldn't hurt to update the kitchen, since I never liked the way long BR's fit in the cans, as shallow as they're set.

I didn't install them and have never been into that ceiling, but I suspect the floor joists above the kitchen ceiling may be only 2 x 6 inch, versus modern 1.5 x 9.25 inch, hence the shallow set to the older cans.

I guess I should buy a few spare? What do folks do when they have 20 matched of the same retrofit, and one goes bad? Surely the same model won't be available for replacement in 5 years? 10 years? One thing about bulbs is that identical replacements have been available for most of the last 60~80 years, which is the blink of an eye to any owner of an old house.
 
Okay, I'm sold. Just started searching, and while I didn't see exactly what I want, already I can see the landscape is way better in retrofits than BR bulbs. Wouldn't hurt to update the kitchen, since I never liked the way long BR's fit in the cans, as shallow as they're set.

I didn't install them and have never been into that ceiling, but I suspect the floor joists above the kitchen ceiling may be only 2 x 6 inch, versus modern 1.5 x 9.25 inch, hence the shallow set to the older cans.

I guess I should buy a few spare? What do folks do when they have 20 matched of the same retrofit, and one goes bad? Surely the same model won't be available for replacement in 5 years? 10 years? One thing about bulbs is that identical replacements have been available for most of the last 60~80 years, which is the blink of an eye to any owner of an old house.
I'd buy at least 2 or 25% extra (whichever is more ) per style as spares.

And if they're on 16 hours a day ( 5000 hrs/ year ) then they will need to be replaced in 7-10 years anyway. Dimmables are a lot more spendy (but usually better engineered and less likely to 'flicker' at 120Hz) and finding a GOOD LED dimmer to drive them can also take an iteration.

But at that usage, they will pay for themselves many times over, even at PA electric rates.
 
Last edited:
  • Like
Reactions: Ashful and begreen
Yes, the rate with which I still am unable to find the same model bulb (not to replace; not needed yet, but to expand lighting with the very same brightness/color etc.) is a bit frustrating.
I understand that happened 15 years ago. The turn over in models these days is annoying.

So buy spares indeed.
 
  • Like
Reactions: Ashful
Look good but $50 a pop seems pretty pricey.
It's $50 for a 6-pack, around $8/ea

They look nice, begreen.

I've never minded the old Halo stuff, when set right, but our kitchen has them set so shallow they just look goofy. No one should have ever put 6" trims in such a shallow ceiling, they should've been 4" maximum. Plus, the Halos, no matter how trendy and sleek they were in 1990, are dated now.
 
$48 for six or $8 each.
Oh i missed that. That seems like a really good deal Now!

I have 1 or 2 that need these. I'll look for the single packs. Thanks
 
Found these at 3000k for 9W, 12W, 15W, and 18W options $9-$16. Plus delivery.

 
  • Like
Reactions: Ashful
All interesting, but completely irrelevant to the failures mentioned above, which were candelabra bulbs in lamp posts, not reflectors in cans. With a total wattage roughly 14% - 15% of the bulb for which the fixture was designed, and power dissipation far below that due to difference in net efficiency, claiming the failures were due to heat is a tough sell.

I understand the problems well, I spent most of my career cooling high power transistors and associated bias circuitry. I was on the high-dollar low-volume side of the business, where we did not have to worry nearly so much about budgetary constraints, so our solutions were very different. But the underlying problems were the same.
I think you missed the comment you had made that I responded to - it was from a discussion in March. Near that you discussed:
My location is definitely not wet. This pair is tucked way up inside a pair of large PAR38 cans in a porch ceiling, with at least 6 feet over overhang and a dropped soffit.

So I was out of your current loop discussion. Back then you were discussing cans. The email notification I got this week linked to that area of this thread.

So my comments aren't related to candelabra in lamp posts (though that may be a significant temperature issue). In general, being a discerning lighting customer means you have to look deeper into the high quality systems (almost commercial for most items ...).

Enjoy your lighting, however you set it up.
 
  • Like
Reactions: Ashful